A key gap in the body of knowledge contained in both the first edition and the forthcoming second edition of the Highway Safety Manual (HSM) is the limited consideration of traffic speed effects in the predictive models for most facility types. It is well-established that crash severity increases with vehicle speed, and speed may also influence crash frequency. However, because speed is correlated with nearly every other variable used in HSM predictive methods, traditional regression models do not effectively capture speed relationships and may even produce counterintuitive results.

Improving how speed effects are incorporated into these models will result in more realistic and actionable outputs. Research is needed to identify effective ways to include traffic speed in HSM predictive models, thereby increasing their accuracy and alignment with the Safe System Approach. Given the strong correlation between speed and various roadway and intersection characteristics, alternative approaches to quantifying speed effects—beyond standard regression modeling—should be explored.

The objective of this research is to formulate and develop predictive methods that realistically account for the effects of traffic speed on both crash frequency and severity, with the aim of inclusion in a future edition of the HSM. The research will involve a comprehensive review of prior studies on speed-safety relationships (from both U.S. and international sources), evaluation of alternative modeling approaches, quantification and comparison of model results, and development of predictive methods using the most promising strategies.

Ultimately, this research will support transportation agencies in their efforts to eliminate fatalities and serious injuries, many of which involve excessive speed, and will aid in the development of context-sensitive roadway designs that promote safe operating speeds.